JP3973481B2 - Echo canceller system - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an acoustic echo canceller system applied to a hands-free device and a video conference system.
[0002]
[Prior art]
Conventionally, in an acoustic echo canceller system, as a technique for reducing the amount of echo, as disclosed in Japanese Patent No. 0281888, Japanese Patent Application Laid-Open No. 2000-59270 and Japanese Patent Application Laid-Open No. 2000-040985, an input / output gain is set. What is adjusted is common.
FIG. 5 is a block diagram of a conventional echo canceller system. This echo canceller system performs processing for reducing a residual echo, a microphone 1, a volume control unit 2, a subtractor 3 that subtracts the calculated estimated echo and the actual echo, a volume control unit 4 that adjusts a gain sent to the line, The noise control unit 5, the transfer function between the microphone 1 and the speaker is estimated, the adaptive filter 6 having a function of automatically updating the filter coefficient, the speaker 7 for transmitting voice from the line, and the volume of the speaker 7 are adjusted. A volume control unit 8 and a volume control unit 10 for adjusting the gain of sound from the line are provided.
When optimal gain adjustment is performed in each of the volume control blocks 2, 4, 8, and 10 in the figure, the amount of echo transmitted to the line is minimized. Another means for reducing the amount of echo is to use a directional microphone.
[0003]
[Problems to be solved by the invention]
However, when the gain is adjusted, the amount of echo is reduced, but in actual use, there is a problem that the sound is small and the user is difficult to hear. In particular, if the gain is lowered by the volume control unit 10 shown in FIG. 5, it becomes difficult for the user A to hear, and if the gain is lowered by the volume control unit 4, the user B becomes difficult to hear. As described above, since the adjustment of the echo amount and the user's ease of listening are in conflict, there is a limit to the gain adjustment alone.
Therefore, the adjustment of the amount of echo is considered to remove not only the gain but also unnecessary sounds (audible masked sounds) that cannot be heard audibly. The sound generated from the speaker on the user A side is affected by the transfer characteristics on the user A side, so that a sound that should not be heard (hearing masked) may be input to the microphone.
An object of the present invention is to provide an echo canceller system capable of removing such echo components and reducing the residual echo without causing discomfort to the user.
[0004]
[Means for Solving the Problems]
In order to achieve the above object, the invention according to claim 1 generates an estimated echo signal by referring to an audio signal from a line, and estimates a transfer function between a microphone and a speaker from an actual residual echo signal. An echo canceller system having an adaptive filter having a function of updating a filter coefficient and a subtractor for subtracting an estimated echo signal from an actual echo signal input from a microphone , and an audio signal transmitted from a line And a masked component removing unit that removes a masked component in a predetermined auditory characteristic of the voice signal gain-adjusted by the voice control unit. The sound signal from which the masked component is removed is used as sound output from a speaker . The invention according to claim 2 is the echo canceller system according to claim 1, wherein the masked component to be removed by the masked component removing means includes simultaneous masking, forward masking, and backward masking. The component is masked in any of the auditory characteristics.
[0005]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a configuration diagram of an echo canceller system according to an embodiment of the present invention. This echo canceller system performs processing for reducing a residual echo, a microphone 1, a volume control unit 2, a subtractor 3 that subtracts the calculated estimated echo and the actual echo, a volume control unit 4 that adjusts a gain sent to the line, The noise control unit 5, the transfer function between the microphone 1 and the speaker is estimated, the adaptive filter 6 having a function of automatically updating the filter coefficient, the speaker 7 for transmitting voice from the line, and the volume of the speaker 7 are adjusted. A volume control unit 8, an auditory mask processing unit 9 that removes unnecessary sounds (audible masked sounds) that cannot be heard audibly, and a volume control unit 10 that adjusts the gain of the sound from the line are provided.
First, a first echo cancellation processing example will be described. The sound on the user B side transmitted from the line is first gain-adjusted by the volume control unit 10. Thereafter, a component determined by the auditory mask processing unit 9 to be inaudible to the user A is removed. Here, simultaneous masking is used as a model to be removed. The simultaneous masking is a model in which the acoustic component to be masked and the acoustic component to be masked exist in the same region as shown in FIG.
The signal processed by the auditory mask processing unit 9 becomes a reference signal for the adaptive filter 6, is transmitted to the user A side, is adjusted in volume by the volume control unit 8, and is transmitted to the user A by the speaker 7. The sound generated from the speaker 7 is input from the microphone 1 as a real echo.
Here, since this signal has a component that is determined to be inaudible by the user A in the auditory mask processing unit 9, the component can be detected by the user B under the influence of the transfer characteristics on the user A side. It is not input as a real echo. The signal from the microphone 1 is then adjusted in gain by the volume control unit 2 and subtracted from the output of the adaptive filter 6 by the subtractor 3. The subtracted signal becomes a residual echo, is subjected to gain adjustment by the volume control unit 4, passes through the noise control unit 5 that removes noise, and is transmitted to the line.
Next, a second example of echo cancellation processing will be described. The sound on the user B side transmitted from the line is first gain-adjusted by the volume control unit 10. Thereafter, a component determined by the auditory mask processing unit 9 to be inaudible to the user A is removed. Here, continuous masking is used as a model to be removed. Successive masking is a model in which the acoustic component to be masked and the acoustic component to be masked are temporally different as shown in FIG.
The case shown in FIG. 3 is when the sound to be masked is before the sound to be masked, which is called forward masking (forward masking). Further, as shown in FIG. 4, the masking sound may be after the masked sound. This is called retrospective masking (reverse masking).
The auditory mask processing unit 9 removes these components. The signal processed by the auditory mask processing unit 9 becomes a reference signal for the adaptive filter 6, is transmitted to the user A side, is adjusted in volume by the volume control unit 8, and is transmitted to the user A by the speaker 7. The sound generated from the speaker 7 is input from the microphone 1 as a real echo.
Here, since this signal has a component that is determined to be inaudible by the user A in the auditory mask processing unit 9, the component can be detected by the user B under the influence of the transfer characteristics on the user A side. It is not input as a real echo. The signal from the microphone 1 is then adjusted in gain by the volume control unit 2 and subtracted from the output of the adaptive filter 6 by the subtractor 3. The subtracted signal becomes a residual echo, is subjected to gain adjustment by the volume control unit 4, passes through the noise control unit 5 that removes noise, and is transmitted to the line.
[0006]
【The invention's effect】
As is apparent from the above description, the present invention has the following effects.
[0007]
In the echo canceller system according to the first aspect, the adjustment of the gain alone is a limit, but by using the auditory characteristic, it becomes possible to reduce the echo amount without giving a burden to the user. Can reduce residual echo.
In the echo canceller system according to the second aspect, by using the successive masking, it is possible to further reduce the amount of echoes without giving a burden to the user, and it is possible to reduce the residual echo.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of an echo canceller system according to an embodiment of the present invention.
FIG. 2 is an explanatory diagram of simultaneous masking.
FIG. 3 is an explanatory view of joint masking (forward direction).
FIG. 4 is an explanatory diagram of joint masking (retroversion).
FIG. 5 is a configuration diagram of a conventional echo canceller system.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Microphone 2 Volume control part 3 Subtractor 4 Volume control part 5 Noise control part 6 Adaptive filter 7 Speaker 8 Volume control part 9 Auditory mask process part (auditory mask process means)
10 Volume controller

Claims (2)

An adaptive filter with a function to generate an estimated echo signal by referring to the audio signal from the line and to update the filter coefficient by estimating the transfer function between the microphone and the speaker from the actual residual echo signal, and input from the microphone An echo canceller system having a subtractor for subtracting the estimated echo signal from the actual echo signal,
A voice control unit for adjusting the gain of the voice signal transmitted from the line;
A masked component removing unit that removes a masked component in a predetermined auditory characteristic with respect to the voice signal whose gain is adjusted by the voice control unit ;
The echo canceller system , wherein the audio signal from which the masked component has been removed by the masked component removal means is used as a sound output from the speaker .   2. The echo canceller system according to claim 1, wherein the masked component to be removed by the masked component removing means is masked in an auditory characteristic of any one of simultaneous masking, forward masking, and backward masking. Echo canceller system characterized by being a component.

Images